Laserfiche WebLink
490 S.M: NELSON AND D.C. ANDERSEN <br />iterative method but correcting for the first and succeeding axes. Rare taxa (those occurring on only a <br />single occasion) were deleted and data log transformed [ln (X+ 1)] before analysis. To test the validity of <br />the CA pattern, the original data-set was randomly rearranged (n = 10 matrices)- to create a null <br />hypothesis assemblage. If the amount of information contained in the original data-set differs significantly <br />(p 5 0.05, 1-tailed test) from the random data-set, it suggests that the pattern is meaningful (Bloom,-1994). <br />Nectar resources and canopy shading <br />During the 1997 field season, the number of flowers or inflorescences considered nectar sources were <br />estimated during butterfly surveys. Although these counts are not a direct measure of nectar, a linear <br />relationship between amount of nectar and number of inflorescences has been found, and it has been <br />suggested that the effort needed to quantify sugar production adds little additional information (Holl, <br />1995). Sampling took place within a 5 m radius circle at disjunct locations every 15 min during a butterfly <br />survey; 16 samples were collected per site each survey month. One-way ANOVA followed by Tukey's test <br />was used to test for differences in mean transformed [ln (X+ 1)] values between sites during each survey <br />month. <br />We used Lux readings from a handheld light meter (Extech®) to compare shading at the sites. We used <br />the percent of the light value under canopy relative to the value in the open as an index of shading at each <br />site. ~ <br />RESULTS AND DISCUSSION <br />Species richness <br />Both species. richness and total butterfly abundance fluctuated seasonally. Mean species richness (Figure --- <br />3a) tended to :"peak in -June regardless of site age, with highest values .consistently. recorded at BW. <br />Abundance values generally paralleled species richness except for the tamarisk-and some revegetated sites <br />in August; when large numbers of Colias eurytheme [associated with -offsite fields of alfalfa.- (Medicago <br />saliva), a host plant] were present (Figure 3b). All sites showed a marked decline in species richness in <br />1996 relative to earlier years, perhaps reflecting a regionwide drought during winter 1995-1996 (Figure 4). <br />Taxa richness, for example,. from combined March and June samples at Bill Williams was 21, 24, and 15 <br />for 1994, 1995, and 1997, respectively, and only 9 for 1996. Because of consistently low values, we did not <br />include the 1996 richness data in the set used to evaluate the relationship between species richness and site <br />age. <br />We found no evidence that older revegetated sites support a greater richness of butterflies than younger <br />sites. In fact, the correlation between butterfly species richness and site age was negative (Spearman rank <br />correlation coefficient, r = - 0.2483; Figure 5). A statistically significant positive correlation (p < 0.05, <br />1-tailed test) would have been reached only with a coefficient value, r > 0.484 (n = 13) (from Zar, 1984). <br />The lowest richness value for BW (Figure 5) exceeded the highest value found at any other site. Richness <br />values at revegetated sites and the tamarisk site were similar, perhaps because a small amount of mesquite <br />and quailbrush was present at the tamarisk site, a trait typical of tamarisk-dominated areas along the <br />lower Colorado River. A species list by site is presented in Appendix A. <br />Nectar resources <br />Revegetated sites tended to have fewer nectar sources than either BW or Tam (Figure 6), Nectar <br />sources, however, were occasionally numerous at moist Havasu Refuge sites. Nectar sources were most <br />numerous in March and June, and rarest in August and November (Figure 6). Nectar sources were low <br />at all sites in November. Although nectar sources were abundant at both Tam and BW, different types <br />of flowers were involved (Figure 7). Most of the flowers or florets at Tam were from tamarisk, while <br />seepwillow flowers contributed more at BW. Other common nectar plants at the sites included mesquite, <br />Copyright ©1999 John Wiley & Sons, Ltd. Regut. Rivers: Res. Mgml. 15: 485-504 (1999) <br />r <br />